You are on page 1of 9
i Lior <4) ~ 173-03 2e ®_ Reassessment of Offshore Steel Structures: Shakedown and Cyclic Nonlinear FEM Analyses @yvind Hellan*, Bjorn Skallerud*, Jorgen Amdahi** and Torgeir Moan** + SINTER Strctural Engineering, N-7034 Trondheim, Norway “+ Division of Marine Structures, ‘The Norwegian Institute of Technology, N-7034 Trondhéim, Norway eraopuerzoy Offshore structures are generally designed according to the Ultinate Limit State (UL), the Fatigue Limit state (PLS) and the Serviceability Limit State (SUS) criteria. TRe conventional UES check is based on Linear elasiie analyses. Each monber 15. checked with cespect to yielding, buckling etc., for environsental loads with a retumn period of typically 100 year, and for specified functional Loads. Some codes (ECCS, 1978 and The Norweylan Petro: teun Directorate, 1984) have introduced deziga cri toria to ensure gurvival of the structure in case of accidental loads of abnoreai strength, typically with ® return period of 10 000 years. ‘These checks are carried out in the limit state of Progressive Collapse (Pts). tn the PLS control, elastoplastic behaviour is, accepted as long as the overall integrity of the Structure is maintained ‘The relationship between ULS and PLE analyses is illustrated in Figure 1, for vave loading. Iasteed of applying strictly the 10 000 yea wave Load, it has been common practice to increment the 100 year vave forces proportionally up to collapse of the structure. This is partly because the reserve strength as such has received considerable interest. and partly because the 10 000 year wave load seldon governs the design Such an analysis is generally called a static pushover analysis: Collapse analyses of the whole structure have documented a significant reserve strength, with system collapse occuring a leads often 50% shove the load that cause first menber failure. The question has been raised as to vhether these reserves can be of any benefit’ - and how mich of the reserves can be utilized, ‘The design conditions for a platform may change during the service Life due to © Increased topside (operational) Loading © extended Life beyond the design service Life ‘© Re-appraisal of the environmental Loading © Reduced load-carrying capacity due to danages, Te consequence of these events may be that first. somber Failure occurs for an environmental Load smaller than the revised design load. If the ULS Gesign ceiterion is to be met at any tine ducing operation in the conventional way, the structure may either have to be strengthened, or the functional Joads mist be reduced. This may be a very expensive task. On the other hand, it is unlikely that the structure will collapse for a single design wave action, due to the inhereat strength reserves that can be demonstrated In 1988, Shell published a progressive procedure for re-assessment (Stewart et al, 1988) utilizing non- Linear static pushover analyses for structures vnoze integrity could not be denonstrated using conventional Linear methods. This paper suggests how these more common pushover analyses can be extended : Instead of analyzing the structure under single wave action, the structure is subjected to a series of waves, to account for the cyolic nature of the loading. A methodology, complesenting the previous push- over procedures, is outlined for integrity assessment of structures under revised “design” conditions. The nnethodology is based on the following simple observations © most offshore structures have a significant reserve strength, and a significant potential for redistri- bution of internal forces, © tie dowinating Ioading on offshore structure is a repeated, variable Leading [vave loads). Nonlinear shakedown analyses 1s epployed in order to investigate ir the structure is likely te fait fron the repetitive action of the waves, or if the struc ture reuaing stable ‘mis new approach to reassessment is being in- vestigated within 4 nev project sich is sponsored by Eeveral oil companies and engineering consultants. BASEC CONSIDERATIONS Assune that the design conditions for 9 structure have been changed so that the ULS criterion (first ember failure) is no longer satisfied outing the ronaining Lite, tne structure will be exposed to a large nunber of waves, Ae illustrated in Figuce 1 the vast majority of the vaves produce only elastic response. Hovaver, sone few, Large waves may cause elssto-plactic responce; i.e. the waximun wave in the remaining Life (design wave}, and depending on the load/strength “mismatch”, the second highest wave, the third highest vave ete Prog wave statistics, it is shovn (Gumbel, 1958) that in the order of $ to’ 20 waves are Likely to exceed 90% and 90% of the design wave haight, respec tively. Hence, the number of waves to be considered that cause plastic deformation, is small. In order to Clarify the following discussion, the tise history has been ordered according to decreasing size in Figure 2 For simplicity it may be assumed that the strue~ ture is subjected to vaves with the sane magnitude as the maxinun wave/design wave. further, it is shown ia Luyties and Geyer (1987) and Wellan et al (1990) that the maximue load in the opposite direction during passage of the wave is analler than the design wave Toad, in the order of 25-408. nen the structure is loaded up to the design toad level, yielding takes place in the sections whece the vis criterion is violated, Sedistcibution of forces may also lead to yielding in other sections in the structure. As shown in Figure 3, yielding causes a reduced overall stiffness, and produces permanent plastic deformations when the structure is unloaded Nhen the structure is celoaded, yielding may take place once again. Within a specific load range, rope> Citive actions of the load give smaller and smaller Ancrenents io the plastic deformations. tn the end the structural Benaviour may be fully elastic, but with permanent plastic deformations in sone parts of the structure, ar illustrated a Figuce 3. The perma- nent plastic deformations sn the initial cycles induce a residual streee {ipla in the stzuctuce. In the Subsequent cycles, thie stress field counteracts tho fect of the wave load, and the structure reaches shakedown. On the other hand, if the Load/strength “mis atch" is higher, the plastic deformation increments fay not decrease during the lead cycles. In this case, the accuatlated displacements sooner or Later grow 30 lacge that the structure becomes unstable and fails aue to incremental collanse (Pigure 4) ‘these two situations are also schematically iliy- strated In Figure §. ‘The leading is represented by a total load rosette or @ load donain 2. There exists a maximus size of the donain corresponding to shakedows of the structure, For loads outside thie donain, the structure will fail due to incremental collapse. So far, ductile material behaviour has been accused. Novever, fracture due to alternating plasti- city (low eyele fatigue) ox extensive inelastic ateaining of the material ay occur. Tt does not Tepresent global failvee ar such, but may prevent the Structure fron eeaching the shakedown state, or may accelerate the inceamantal collapse process (Figure 5 SHAKEDOWN ‘THEORY ‘the following questions should be answered if 3 structure is subjected to variable loads that cause plasticity: = for a given steucture snd load domain 9, will the Structure reach shakedown avoiding incremental Collapse and/or alternating plasticity for any load path contained within this domain? . He the struetuee shakes down, will the resulting deformations be of acceptable agaitudes? = i the steucture shaker dovn for a domain 9, what is the safety againe incremental collapse or alter ating plasticity (1.8. how much may the load Gonain be enlarged)? ‘The shakedown theoxens state that, when ful- fitted, a structure subjected to variable leads vith Unknown history but known load domain is safe with Respect to these failure situations. Tf the structure Shakes down, further variable Loading of the saae (oF Shaler) magnitudes dove only cause #lastic deforma tions, i.e, the plastic defornations have becom Stationary. Coresponding te this plastse field, a Statiouary residual stress field p (%, €) = 2 1X tunrssl has also developed. Stationary fields may be eumally expressed as (Stein et al, 1990) aie OF ie, t= 0 a Lim p(x, t= 0 @ nese relations express whether shakedown ocevr Also, the accumulated plastic work in the stcuctuce ust be bounded (Kénig, 1987), or else incremental Collapee/atternating plasticity occur ® wef J o:eacavce oy va physically, the ability of saterials and strue- tures to accumulate plastic work is Limited due to exhaustion of ductiLity/low cycle fatigue, vhers Significant cracks have developed and sudden fractures hay occur. Hence, the infinite term should be replaced by some Finite, physically based, magnitude Wpinex- Te the structuee has reached the shakedown state for an arbitrary Load path within the load domain, the yield criterion, £q. (4), is satiefied everywhere in the structure. ‘This means that the external loads are Carried elastically, and the stresses in the structure nay be written as the sua of an elastic field and the residual stress field, Eq. (6) elove, 8) $09 OD ) Pet oj $0 ca) oro sp co or tof cx, ©) © POAT £09 Jinece of denotes the elastic stresees and p denotes the cesidual stress field caused ny pernanent plastic Getormations. 5 is the yield stress Now the necessary and sufficient static shakedown theorem states that (Melan, 1938) A) If there exiete a time-independent residual ateose field p(x) anda factor a> 1, sucn that tem (of ix, ths BRINE Et ts satisfied for all loads defaning 9, and Cor all x in the structure, tien the structure will chake doun for the given load domain Note that 6 may be different from the exact p. the 1oad domain, Figure 6, may also be described by the verticus (vectors from the ozigin to the corners of the hypec-parallellepiped 2}. The folloving theorem decived by Xnig and Rleiber (1978) nay be applied in order to discretize the time variation b) Kf a given structure shakes down over @ cyclic load path containing all the vertices D* of the Load Gonain, then it shakes dova for any load path contained within a A consequence of this theorem is that the chako: Gown process may be investigated by cycling sequent ally along ali the vertices, where each vertice corre: Sponds to @ proportional on- and eff-loading of the Steucture. Beth theorens presute linear geowetry and convex yield eurfaces Teaditionally optimization techniques have been applied in solving shakadoun probloas tm such methods the unknowns 9 and maximum size of the load domain that setisfles A) are searched for. Hence the theoren is directly applied as a consteaint in the golution (Stein et al, 1990). One disadvantage With this approach is that no information of the accu Solated plastic deformations is obtained. sont inear Geometry also coplicatos such an analysis Another approach is based on incremental on- Linear PEN analyses. Presently, nonlinear structural anclyses with proportiosaliy increasing Loads ace perforned fre. Guently due to the availability of efficient PEM. Godes. the response is calculated by stepwise solution Of the equilibrium equation on incremental form. e.g oe - ey (ride « Bre Re +R o were the tangential system stiftness matcsx Ky say account for nontineer material behaviour (2-5 plasticity) and nonlinear geometry (e.g, stabilising/ destabilieing effects Like °-5) tn analysing shakedown by means of the incr’ ental approach, theores 8} may be applied. Now the Fesponse is computed for sequential cycles of propor Uieral loading along the vertices describing the Load Gonain, monitoring the plastic strains. Usvally, the funber of cycles required to determine if shakedown Securs, io relatively small. The plastic deformations aR okt ae em ann nett amen are direct output fron the FEM analysis. If the sneze- Rents of plastic deformations approach zero during the Toad cycling, the structure shakes down for the given load domain and the yiela cesterion ie eatistiad avecywhere in the structure (in the integration pointe). Hence, if the situation described by Sas (1) 3nd (2) obtained, and theorem a) is fullfilled to determine tho maximum size of the shakedown domain new cyclic PEN anelyses are performed until, for sone size of the toad dowain, shakedovn does not Occur. ‘The previous domain may then be employed as 4 neasure for the shaxedown capacity Since the cesulting plastic deformations at shakedown are determined in the FEM approach, they aay Gurectly be assessed against any stcuctural setvice- ability requirements. One may also calculate accumu- fated plastic work or etraine that may be conpared with sone physical fracture eriterion, @.g. besed on a Tepresentative low cycle fatigue capacity curve. in this approach, geonetrical noplinearity may also be accounted for ina Airect manner. But some work Tonains in order to determine whether there are sequ- nce effects fron’different choices of cyclic patterns fon the conputed maximus values of the load domain. Some sequence effects should be expected since ditt Pent eyelic patterns yield different plastic deforsa- Clon histocies, hence also different P-6 evolutions. Solution of shakedown probless by nonlinear FEM codes is a relatively novel appcoach. Borkowski and Kleiber (1980) analysed shakedown of truss works with finple axial elenente aad 2 cantilever with Linear Steain triangle elenents. Linear geometry vas asauned in these analyses. Nonlinear shakedown of beans and Frones analysed by the notlinear FEX are performed by Skeliered (1986) aca Skallerud and Larsen (1969). Here the US*0S computer code described subsequently was applied (Soreide et al, 1900). These analyses showed that nonlinear geonetey was 9 significant effect to be accounted for in calculation of shakedown Loads: FEN FORMULATION ‘the basic idea behind the UsFOS (Soreide et al, 1906 ana 1988) program is to use only one finite Glenent per physical maaber. Thus, the finite eleaent hesh of conventional Linear anelyzis may be used, Hlashicelenent ‘the basic elenent in USFOS is the 3 Aimensional, 12 bof beam element. Using Green strain and von Kacnan theory, the strain displacenent expression on Local slenent level takes the fore grease : cys ut beet ebm 19) where v(x) denotes the axial displacement and v(x) ané Win) ere the Lateral deflections. This expression is valid for Lavge lateral displacements and moderate rotations ‘the total and incremental equilibrium equations aro established by taking the first and second varia~ tion of the strain energy and the potential of external loads (Soreide ot al, 1988) ‘the interpolation polynomial for the axial and lateral displacenents are the exact solution of the £ferential equation for 4 bean subjected to axial force and laterial bending loads at tho bean ends Vix] + Ay con ke sag sin kx aye Ag) Iconpeesesen) UR) = ay eo ay oH Agee Re ay (tension) vere k + gi> and ois positive in compression: Ay ~ Ay ave generalizes coordinates ‘the stiffness matrices for the nonlinear elastic bean elasent is derived ae closed form solutions. 0 puaeeical integration over element Length of over clement cross section is necessary slasto-plastic element Material nontinearity 1s modelled by plastic hingas introduced at monber ends or at member midepan ‘The yield criterion is expressed in terms of stress resultante (axial ores, shear force, torsion and bending moment). TAN, Gye ee Hae My Mad =o on T+ 0 defines @ plastic state of forces while 1 Sf © 0 gaflects elastic cross sections. = 0 represents a surface in the stress resultant space According to plasticity theory, this suctace is couvex, which Le a fundaneatal reguireaent for shake~ down analyses (Skellerud, 1991) Stiffness qatrices for elements vith plastic hinges ace modified according to plastic flow theory ‘the consistency criterion states that tne state of forces Ines on the interaction surface ducing plastic deforsation, This is expressed for an elastic “perfectly plastic Mage as Aegis 20 (it, 2) on where OS, is the increment in forces and gy! is the Gradient of the yield suztace at each node of the ‘the elasto-plastic stittness aatrices for the elements can now be derived as Key + Ke = hes (aTegg"! oT ky asp whece ky As the elastic, tongential stiffness nate ix Plastic hinges can be introduced at element ends and at midspan. Both gradual plastification and kine matic nacdening of the plastic hinges may be eccounked for in addition to Local buckling, joint flexibility and fracture UsrOS Le based on an updated Lagrange formula with a simple Fuler-Couchy increnentation proce- Gure and local element reference syatens. Nodal co- ordinates are updated after each load increnent, thos aecaunting for large displacements and the *P-8" EXAMPLE STUDIES two-storey Frame In Figuee 7 two-story frame vith horisontal and yeetical Leads varying between zero and a naximun is Gepicted, The nenbere have constant tubular cross~ Section. In aééstion to the variable loads A, and dy, 2 pair of static exial Loads are present in the coluans. In Figure 8 some results are plotted. The vari- able load factor } is plotted on the ordinate axis, Gnd the horisontal displacesent of the upper been is plotted along the abscisea-axis. In the first case an {nerenentel collapse mechanism is capidly developing Inthe second case, vhore the maximun load level is Eeduced from 6.5 to 5, g€411 an incremental collapse Situation is eceuceing. Finally, for a slight redve~ tion of toad level, shakedown of the frame is obtained, Hence cesponds the structure elastically after some elasto-plastic cycles for this load pattern. ‘The necessity of including nonlinearities in the shakedown analysis was pointed out in (Skallerd and Largen, 19991 Plane Jacket structuce A plane model of @ jacket with a haight of 184.5 fn is shown in Piguce 9 (Stewart et al, 1988). Typical Inonber diameters are 0.9 - 1.08 for the horizontals tnd 1.3m for the races, Morizontals and braces of the upper panel are 0.6m. Leg diameters range from 13 mat the top to 6.0 mat mudline. D/t ratios vary flon'26 to $7 foe the horszontals, «2 to 66 for the Giegonale and 41 0 121 for the legs: the finite element model consists of 24 nodes and 47 elements, Each member is modelled by one finite Clement. Initial deformations are not considered, and Joints are nodelled as rigid. Local failure modes of the braces and joints aze disregarded, but in-plane and out-of-plane defarmations are considered. The Structuce is subjected to @ topside load of 50 MN and 2 total horizontal load of 46.7 HW. two diffecent cases are studied ~ each represent ing 2 situation where the design conditions have Changed, thus cequicing @ reassassmant of the struc~ ture (Helian, 1990) case 1 Increased environmental Loading Case 2 Danaged/cenoved structural menbers In each case, a pushover analysis is run in order to determine first ember failure and ultimate capa- Sity of the structure. Subsequently, a cyclic load is introduced, varying between @ tines the design Load, and -0.3°0 Static pushover analysis gives @ collapse Load of 1.79 tines the design load. Plrst yield occurs at 1.24 times the enviconmental load, igure 9b. ‘Ancteased envixonnental Loading the structure is analysed under cyclic Loading with a loading factor = 1.6. Global deformations are shown in Figure 10c. Arter the initial Loading, the increments in plastic Getormation are relatively constant. The accumulated plastic displacement are slightly reduced (the incre fneatal plastic displacanents has opposite sign of the iateaal displacenent) ‘The resvite of this analysis indicate that the structure does not shake dow when subjected to & Gyelic Load varying between 1.6 and -0.48 of the assign toad, A three-nange mechanisa Corns in wesber 36 noth in conpression and in tension. It is probable that Low eyele fatigue will eccur after a relatively small nusber of cycles Next, factor @ = 1.5 the analysis wae then repeated with 2 load Again, shakedown does aot occur, Finally, the structure is analysed with a loading factor of @ = 1.4 ‘the global deformations are shown in Figure 1b ‘The untial Toading results in a small pecmanant de feematian. In the subsequent cycles, this deformation inereases, but the increnents aa plastic deformation Geccease. Eventually, the plastic suerenents vanish, fend the structural response becones purely elastic thas indicate that shakedown take place for a cyclic Load varying between 1.4 ané 0.42 of the aesign load Nomber 26 formes a three-hinge mechanism and undergoes large rotations in the suitial cycle. In & complete cyclic analysia, this elements should be checked with respoect to Local failures Aunoved stiuctucal menbers Nonber 26 renoved is renoved from tho model, Figure 9. The structure is analysed with # loadica factor + 1.0 First yield seus at 81% of the design load. The collapse load is 1.59 tines the design Load ‘the global deformations axe shown in Fagure 11. the initial Loading gives @ sali permanent defarna tion, The accumulated plastic deformation increases Ln the subsequent cycles. but the increwents decrease in the Final cycles, tne plastic increments vanish ‘hie andicater that the structure reaches the shakedown atate with momber 36 removed, when subjected too load varying between 1.0 and -0.3 of the design toad LE both elements 36 and 37 are removed, a situa tion of alternating plasticity develops, as shown in ASSESSMENT PROCEDURES ‘tne preceading examples illustrate the potential of cyclic elasto-plastic analyses. However, a nunber of points need to be discussed before such an aporoach Piret of all, the loading magnitude must be de- fined, Here, ono approach would be to apply the most probable distribution of extrone waves. A nore conser Wative alternative would be to apply the design load in all cycles. En the first case, possible sequence effects must also be investigated Next, appropriate critera describing structural capacity needs to be defined. A conservative approach Would be to require shakedown of the structure. A nore refined criterion could be to accept an incromontal collapse process to develop, as long as final collapse fdoee not accur within the design Life of the struc~ Tt is of course recognizes that the shakedown approach may roguire rather Lengthy simulations, and easily Becomes prohibitive vith respect to computer tine consusption for complex steuctures. A najor cbyective of the xeseorch project is theratore to Establish some simplified method to reduce the need for complete cyclic analyses. It would be benificial if the method could e.g. be based on the reserve strength factor (Lloyé and Clawson, 1983) and the Utilization of the structure, since these factors are eetablished fron ordinary pushover analyses. Ultimate strength (systen Reserve strength = as seeeat tization + oa esa tend ——__ Gin Design strength (component) Quantitative information for acceptable utiLiza~ tion/reserve strength vill be provided through the on~ going research project. Extensive cyclic analyses of Sctuel jacket structures will be carried out to in~ vestigate the influence of different layout, bracing Systens, senber compactness, D/t ratios etc. on the cyclic strength of offshore structures, ond to in- vestigate the correlation between cyclic strength and static pushover strength for different classes of stroctures ‘the principles of a possible acceptance criterion is illvetrated in Table 1 Possible acceptance criteria for elasto-plastic Le design Table 1 [CONCLUDING REMARKS A methodology has been proposed for integrity assosament of structures under revised design condi tions. ‘The methodology i based on the following imple observations: # Nost offshore structures have a significant reserve Strength, and @ significant potential for Fedlstribution of internal forces. © ‘Te dominating loading on offshore structure is 2 repeated, variable loading. ‘The methodology employs shakedown analyses by means of nonlinear FEM analyses in order to investi- Gate if the structure is likely to fail from the Eegptitive action of the vaves, or if the structure will remain stable case studies have been carried out for two plane Iremes. one of which is representative for a jacket Structure, for a defined, cyclic Load history below fcertain range, shakedown takes place, i.e. the strve= tural response Decones elastic for a load Level above first yielding ‘The cases indicate that offshoce structures can be utilized beyond the conventional ULS design linit (Eirst member failure}, end still fulfill the basic regulatory requirements (elastic behaviour) after a uber of eyclas. Autnough at a preliminary stage. the results so for are very encouraging with respect to using 2 syelic approach for justifying acceptance of increased utilization in connection with reassesenent of struc Doplication of the proposed methodology require extensive investigations of cyclic, theee dimensional Denaviour of offshore structures, including possible load direct ion/ioad saquence effects. Further work 1s needed to establish reliable failure criteria for con- ponent behaviour under cyclic ioading in the elasto- Plastic range, concerning lecal buckling, fzacture and low-cyele fatigue. These sspects will be addressed in farther within the new Reassessment project ACKNOWLEDGEMENT ‘The present study has been done as a part of the research project "Reassesenent of Marine Structures sponsored by the following companies + Shell, Norsk Hyaeo, Phillips, Saga, Awoce, Aker Gagineersng and Offshore Design. The authors wish to acknowlege the input of ideas provided from the project sponsors. Tn particular, the conteibution from project chasrmen G.Stewart of Shell Reseach is highly appreciates REFERENCES Borkowski, A. and Kleiber, M. (1980), “Oa a Numerical Appraach to Shakedown Analysiz of Steucturee”, Combs Meth, aval. Nach, nang, Vol. 22, pp. 101-119 ECCS (1978). “European Reconuendat ions for Stee! Con steuction® “Statistics of extremes”. Colum New York Gumbel, £4. (1958) hia Univereity Pre! NeLlon, ., Andahl, J., Farnes, H.-A. and Kerunakaran, D, (1990); “Reaseeeement of Steuctuces, Fundamental Considerations", SINTEE Report SIF71 420022, Trond- MeLlan, ©. (1990), “cycLic Analyses of 20 Jacket ‘Structure’, SINTEE Report STE7! AQ0017, Trondne a, KSalg, JA, (1987), "Shakedown of Elastic-Plastie Structure", Elsevier Publishers Konig, JeA. and Kleiber, M. (1978), “On a new method fof shakedown analysis", Bull. Acad, Fol. Sci, Vol 26, pp. 165-171 Ltoya, JR. and Clawson, W.c. (1983), "Reserve and Residual Strength of Pile Founded Offshore PLat~ forms", Exot. Int, symp. kale Dea. inp. xedungancy. warine’ struct. luyties, Wl. and Geyer, J.P. (1987), "The Development ‘Of Allowable Fatique Stresses in APL RPZAT, ORC Raper $558, Houston, 0Sk Melon, £. (1930), “Der Spannungszustand eines Hases Nenchyschen Kontsnuoms bet vezanderlicher Belast- ung", Akad Wiss. wien Ta, vol. 147, pp. 13-78 The Norwegian Petroleum Orrectorate (1984), Regula. tions for load-carrying structures for extraction er exploitation of petroleum", (in Norwegian), Stavanger Skallorud, B. (1986), "Nonlinear Effects on the Shaki ‘down Load of Svay Franes and Continuous Seams", Div. Stevet. Engng., NTH, Trondheim Skatlerud, 8. and Larsen, P.R. (1989), Effects on shakedown of Sidesway France: Engna., ASCE, Vol. 115, pp. 221-227, Skolierod, 8. (1991), “Reassessment of Structures ‘Shakedown Theory’ and Relevance for Offshore Steel Platforne*, SIMREF Resort S2F11 £91005, Trondhesa Stein, &., thang, G. and Nahken, R. (1990), "Shake: om Analysis for Perfectly Plastic and Kinenatic Nocdening Materials", Stodola Session "Progress in Computational Analysis of Inelastic structures: vaine, ttaly. Stewart, 6., Efthymioy, H. and Yagts, FM. (1988) “Ultimate Strength ang Integrity Assessment of Fixed Fixed Offshore Platforms", Eroc, Fifth BOSS Confer= ance, Trondheia. Goreide, 7-H. et al. (1986), "Collapse analysis of Prawed Offshore Structures", O72 Paner 5202, Houston, 08, ‘ust0s Theory Manual”, 1990-07-01, ‘Trond Soreiae, 7.4. et al. (1908), SINEE Report STF? yAG026, rev. osu, Norway, Figure 7

You might also like